Surf.new

Surf.new is a free, open-source platform that deploys AI agents to automate web browsing and research by mimicking human navigation and interaction on websites. It allows users to automate repetitive tasks such as data scraping, price monitoring, and market research.

AI agents powered by models like Claude 3.7, OpenAI, Gemini 2.0 Flash, and DeepSeek R1 navigate web pages based on user instructions, retrieve data, summarize content, and can perform automated actions. This process simulates a human-like web browsing experience but automated and faster.

Its main features include web navigation and data scraping, automation of tasks like form filling and CAPTCHA solving, supporting multiple AI engines and easy toggling between them, real-time visualization of browser interaction, and customizable AI agent configurations for different research and automation needs.

From developers who want to test AI web agents to individuals or businesses looking to automate online research and browsing tasks, Surf.new offers an accessible experimentation environment with broad applicability.

No, Surf.new is not a traditional browser for consumption but an AI-powered tool designed to automate web interaction and support research or data-driven tasks. It builds on "AI browsers" concepts but focuses on automation and agent experimentation rather than just chatting with an AI assistant.

Surf.new runs web content in isolated sandboxes and supports local data storage for notes/resources. When it uses smart features, data in context is sent to LLMs on their servers, but users can provide their own API keys or use local language models for privacy and control.

Yes, the platform is free and open source, enabling users to get started immediately for research or automation purposes.

It supports multiple large language models and frameworks including Claude 3.7, OpenAI's models, Gemini 2.0 Flash, DeepSeek R1, Steel.dev's AI browser, and LangChain orchestration.

Surf.new is actively evolving as an open-source playground for deploying and testing AI web agents, allowing continuous improvements and feature additions.